Evaluation of electrospun polymer–Fe₃O₄ nanocomposite mats in malachite green adsorption

Abstract

Magnetoactive nanocomposite fibers, based on poly(ethylene oxide) (PEO), poly(L-lactide) (PLLA) and pre-formed oleic acid-coated magnetite nanoparticles (OA·Fe₃O₄), were fabricated by electrospinning and evaluated for the first time as substrates for the adsorption of N-methylated diaminotriphenylmethane dye (malachite green oxalate, MG) from aqueous media. The adsorption of MG onto the fibers was investigated under ambient conditions by means of UV-Vis spectrophotometry as a function of initial dye concentration and solution pH. Equilibrium data for MG adsorption were well-fitted with the Langmuir isotherm model indicating a monolayer adsorption process. The effect of magnetite nanoparticles on the adsorption efficacy has been also demonstrated by performing the aforementioned studies on fibers that did not contain OA·Fe₃O₄. The obtained results suggested that the presence of embedded magnetite nanoparticles reduces the fiber adsorption efficiency to some extent. Moreover, the thermodynamic parameters determined from adsorption experiments carried out at three different temperatures indicated that the adsorption of MG onto the Fe₃O₄-free and the Fe₃O₄-containing fibers is spontaneous and endothermic. Although the presence of Fe₃O₄ within the fibers disfavored somewhat the adsorption process, nevertheless, the incorporation of the magnetic nanoparticles within these materials assisted their recovery from aqueous solutions by means of an externally applied magnetic field. Desorption of MG from the fibers could be realized upon fiber immersion in alcohol solution, thus allowing the regeneration and re-use of the adsorbents that retained the same adsorption efficiency after multiple regeneration cycles. MG adsorption studies performed in urban wastewater samples by using the PEO/PLLA and the PEO/PLLA/OA·Fe₃O₄ fibers as adsorbents, demonstrated the potential use of these materials in real wastewater treatment applications.

---

Introduction

The extensive use of synthetic dyes in manufacturing such as in textile, paper, paint, food, pharmaceutical and cosmetic industries, has led to an increased scientific interest in the field of dye wastewater treatment.^1,2 The toxic and carcinogenic effects of dye-containing wastewater to aqueous ecosystems^3,4 in combination with the dye resistance to biodegradability and their stability to irradiation, heat and oxidizing environments^4,5 has led to the development of different methods for the removal of dyes from wastewater. Among others, adsorption,⁶ oxidation,⁷ precipitation,^6,8 and coagulation⁹ have been employed for the purification of dye-contaminated aqueous solutions with the most promising method being the adsorption process.¹⁰ Although carbon-based materials have been one of the most popular dye adsorbents used in wastewater treatment,^11,12 their high cost and poor regeneration capability prompted the scientific community to focus on the development of cost-effective, re-usable and highly efficient dye sorbents.^13,14

One of the most common organic dyes used in the textile, paper, paint and ink industries is malachite green (N-methylated diaminotriphenylmethane dye) (Fig. 1).¹⁵ Considering the harmful effects of this dye to the aquatic life¹⁶ and humans, significant efforts have been focused on the development of new materials capable of acting as effective adsorbents for this substance from wastewater including hydrogels,¹⁷ fibers,¹⁸ films,¹⁹ and bioadsorbents.^14,20

Fig. 1 Chemical structure of malachite green oxalate: N-methylated diaminotriphenylmethane dye.

Among others, fibrous materials with diameters in the micro- and nanometer range have attracted considerable attention as potential adsorbents for harmful substances from aqueous media due to their high specific surface area and the ease for regeneration and recycling.²¹ Electrospinning has been recognized as a simple, cost-effective and efficient technique for the fabrication of polymer (nano)fibers.^22–24 In addition, its versatility enables the incorporation of a variety of inorganic micro/nano particles within the polymer fibers resulting to the generation of (nano)composite materials in the form of continuous (nano)fibers and (nano)fibrous assemblies.

Due to their high surface to volume and aspect ratios, low density and high pore volume, electrospun (nano)fibers with functional groups can be effectively used to adsorb pollutants via physical or chemical adsorption.^25–31

Electrospun fibers exhibited high adsorption capacities essential for removing heavy metal ions^32–35 and organic dyes.^18,36 Concerning the latter, only a few examples appear so far on the use of electrospun fibers in dye removal from aqueous media.^18,36–40

The use of magnetic NPs in environmental remediation processes involving the removal of organic dyes from aqueous media receives considerable attention nowadays.⁴¹ Previous reports demonstrated the ability of magnetic NP-containing nanocomposites to act as efficient dye adsorbents in aqueous media.^42–45 Moreover, the presence of magnetic NPs in such applications is considered to be highly advantageous since it enables the easy recovery from waste streams by means of an externally applied magnetic field.

The present study deals with the removal of malachite green oxalate (MG) from aqueous solutions using magnetoactive (superparamagnetic) polymer–Fe₃O₄ fibers as adsorbents, prepared by means of the electrospinning technique.⁴⁶ More precisely, fibers based on poly(ethylene oxide) (PEO), poly(L-lactide) (PLLA) and pre-formed oleic acid-coated magnetite nanoparticles (OA·Fe₃O₄) fabricated by electrospinning were evaluated for the first time as substrates for the adsorption of MG from aqueous solutions. To the best of our knowledge, there is no previous literature report concerning the use of electrospun polymer-based magnetoactive nanocomposite fibers for the removal of MG from aqueous media. Experiments on the effect of pH, initial dye concentration and fiber magnetic content have been performed and the adsorption capacity of the fibers has been evaluated in the absence and presence of magnetite nanoparticles.

Experimental section

Materials

The two homopolymers poly(ethylene oxide), (PEO, M̄_n = 600 000 g mol⁻¹), and poly(L-lactide) (PLLA, M̄_n = 99 000 g mol⁻¹) were purchased from Sigma-Aldrich and used as received from the manufacturer. Chloroform (CHCl₃), ethanol ≥99.9% (EtOH), malachite green oxalate (MG, M_w = 927.02 g mol⁻¹), sodium hydroxide pellets (NaOH), and hydrochloric acid (HCl), 37% were purchased from Scharlau.

Adsorbent fabrication

Electrospun fibers based on the commercially available PEO and PLLA were fabricated in the absence and presence of OA·Fe₃O₄ as described in a recent publication.⁴⁶ The OA·Fe₃O₄ nanoparticles were prepared by following the chemical co-precipitation method as described in previous reports.^47–50 Solutions of PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ (solution concentration: 4% w/v; weight percentage proportion of PEO/PLLA: 70/30 respectively, 37 and 50% wt OA·Fe₃O₄ in respect to the total polymer mass) were prepared in chloroform. All electrospinning experiments were performed at room temperature. Equipment included a controlled-flow, four-channel volumetric microdialysis pump (KD Scientific, Model: 789252), syringes with specially connected spinneret needle electrodes, a high-voltage power source (10–50 kV) and a custom-designed, grounded target collector, inside an interlocked Faraday enclosure safety cabinet. Systematic parametric studies were carried out by varying the applied voltage, the needle-to-collector distance, the needle diameter and the flow rate so as to determine the optimum experimental conditions for obtaining fibers.⁴⁶

Adsorbent characterization

The morphological characteristics of the fibers were determined by scanning electron microscopy (SEM) (Vega TS5136LS-Tescan) prior to and after MG adsorption. The samples were dried in a vacuum oven (after immersion in aqueous media) and then gold-sputtered (∼30 nm) (sputtering system K575X Turbo Sputter Coater – Emitech) prior to SEM inspection. Further information on the characterization of the PEO/PLLA/OA·Fe₃O₄ adsorbents including average diameters of the Fe₃O₄ nanoparticles embedded within the fibers determined by TEM can be found in a recent publication of our group.⁴⁶ X-ray powder diffraction patterns were obtained using Rigaku (30 kV, 25 mA) with λ = 1.5405 Å (Cu) in the range of 20–80° and at a scanning rate of 1° min⁻¹. Thermal gravimetric analysis (TGA) measurements were performed on a Q500 TA instrument under nitrogen flow at a heating rate of 10 °C min⁻¹. ATR-FTIR spectra were recorded using a Shimadzu FTIR-NIR Prestige-21 spectrometer bearing an ATR accessory.

The magnetic properties of the nanocomposite fibers were determined at 300 K with a Vibrating Sample Magnetometer (VSM)-Model 880 from ADE technologies USA. Nitrogen adsorption isotherms were measured with the use of a Micrometrics ASAP 2000 automated apparatus at 77 K. The samples were degassed at 293 K for 24 h.

Preparation of aqueous dye solutions

In order to investigate the effect of the initial dye concentration on the adsorption process, a highly concentrated MG stock solution was initially prepared (1 × 10⁻⁴ M) and used for the preparation of MG solutions of various concentrations ranging from 1 × 10⁻⁶ M to 1 × 10⁻⁵ M. The effect of initial pH on the adsorption of MG onto the fibers was determined at different pH values ranging between 3 and 9. The pH values of MG solutions were adjusted by using NaOH or HCl aqueous solutions (0.1–1 M) while the dye concentration remained constant (C₀ = 6 × 10⁻⁶ M). All experiments were performed at room temperature (298 K) and in all cases the total volume of the solution (3 mL) and the mass of the adsorbent (10 mg) were retained the same.

Adsorption studies

Adsorption studies were conducted by means of batch experiments in polyethylene (PE) screw-cap tubes containing a certain amount of fiber (10 mg) immersed in MG solution (3 mL, prepared in de-ionized water) of known concentration (1 × 10⁻⁶ M to 1 × 10⁻⁵ M) for 24 hours. After equilibrium, UV-Vis spectrophotometry (Jasco V-630) was used to determine the residual dye concentration in the supernatant at the MG maximum absorption wavelength (618 nm). The amount of MG adsorbed on the fibers (q_e) and % removal efficiency (% q_e) were calculated using the following equations:⁵¹where q_e (mol g⁻¹) is the amount of dye adsorbed onto the unit amount of the adsorbent, C₀ (mol L⁻¹) is the initial dye concentration C_e (mol L⁻¹) corresponds to the equilibrium concentration of MG in solution, V (L) is the solution volume and W (g) is the adsorbent (fiber) mass.

Kinetic adsorption studies were performed by immersing a specific amount of either PEO/PLLA or PEO/PLLA/OA·Fe₃O₄ fiber (10 mg) in different plastic tubes containing MG solution (3 mL, C₀ = 1 × 10⁻⁵ M (0.0927 g L⁻¹)), and recording the UV-Vis spectra of the supernatant solution at different time intervals.

The effect of pH on the dye removal by the PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ fibers was studied by placing a pre-fixed amount of the adsorbent in the MG solution of known concentration in which pH was adjusted as described above, and recording the UV-Vis spectra of the supernatant solution after 24 hours. The MG concentration was examined at the pH range 3–8, by measuring variations at the λ_max appearing ∼618 nm.

For determining the thermodynamic parameters related to the adsorption process the above-mentioned procedure was repeated at 318 K and 338 K. Pieces of fibers (10 mg) were placed in MG solution of known concentration (1.13 × 10⁻⁵ M) for 24 hours prior to UV-Vis analysis.

Desorption/regeneration and recycling studies

For the regeneration of the fibers, the MG-loaded PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ fibers were removed from aqueous solutions and were immersed in ethanol solution (3 mL). After 5 min the fibers were removed from ethanol and re-immersed in another MG aqueous solution of the same initial concentration (C₀ = 1 × 10⁻⁵ M). The UV-Vis spectrum of the supernatant was recorded after 24 h for determining the amount of the adsorbed dye. The aforementioned adsorption–desorption cycle was repeated three times.

Removal of MG from urban wastewater

MG adsorption studies were also performed in urban wastewater (UWW) samples (secondary treated effluents) collected from the urban wastewater treatment plant, located at the premises of the University of Cyprus. The UWW samples were first passed through a paper filter to remove insoluble particles and impurities. Kinetic adsorption studies were performed by immersing a specific amount of either PEO/PLLA or PEO/PLLA/OA·Fe₃O₄ (37% wt) fiber (10 mg) in polyethylene (PE) screw-cap tubes containing the MG solution (3 mL prepared in UWW, C₀ = 1 × 10⁻⁵ M), and recording the UV-Vis spectra of the supernatant solution at different time intervals for a period of 5 hours.

Results and discussion

Fiber fabrication, morphological and compositional characterization

Microfibers based on PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ were fabricated by electrospinning.⁴⁶ The successful generation of fibers requires the determination of the optimal processing parameters that include the concentration of the polymeric solution (which significantly affects the solution viscosity), the applied voltage, the delivery rate of the solution, the diameter of the needle and the distance between the tip of the needle and the collector. By performing experimental parametric studies for the aforementioned systems, the optimum processing conditions for obtaining homogeneous fibers were needle diameter 16 G, needle-to-collector distance 25 cm, applied voltage 20 KV, and flow rate 4 mL h⁻¹ in the case of the OA·Fe₃O₄-free fibers (PEO/PLLA) and needle diameter 16 G, needle-to-collector distance 25 cm, applied voltage 25 KV, and flow rate 4.5 mL h⁻¹ in the case of the OA·Fe₃O₄-containing fibers (PEO/PLLA/OA·Fe₃O₄).⁴⁶

The morphological characteristics of the fibers, before and after immersion in aqueous solutions, were determined by SEM. Fig. 2 provides the SEM images of the as-prepared PEO/PLLA (Fig. 2(a)) and the PEO/PLLA/OA·Fe₃O₄ electrospun fibers (Fig. 2(b) and (c)) containing 37 and 50% wt OA·Fe₃O₄ respectively. Highly homogeneous, cylindrical fibers with mean diameters of approximately 2 μm were observed both in the absence and presence of OA·Fe₃O₄ nanoparticles. As seen in the images, no significant changes were observed in the morphological characteristics of the fibers upon the addition of the OA·Fe₃O₄. The presence of the highly water-soluble PEO within these materials resulted to morphological changes of the fibers upon hydration as shown in Fig. 3(a) and (b). Those changes were more pronounced in the PEO/PLLA fibers (Fig. 3(a)) in comparison to the PEO/PLLA/OA·Fe₃O₄ analogues (Fig. 3(b)). This phenomenon is probably attributed to the fact that the OA·Fe₃O₄-containing fibers are more hydrophobic in nature, due to the hydrophobicity of the Fe₃O₄ surface at pH values below 8 ⁵² as well as the presence of the hydrophobic oleic acid chains surrounding the Fe₃O₄ nanoparticles. Consequently, fiber hydration is prohibited or at least decelerated in the presence of OA·Fe₃O₄.⁴⁶

Fig. 2 SEM images of PEO/PLLA polymeric fibers (a) and of magnetoactive PEO/PLLA/OA·Fe₃O₄ fibers with 37% (b) and 50% wt (c) magnetic content.

Fig. 3 SEM images of PEO/PLLA fibers (a) and magnetoactive PEO/PLLA/OA·Fe₃O₄ fibers in the presence of 50% wt OA·Fe₃O₄ nanoparticles (b) after being immersed in an aqueous MG solution.

The nanocrystalline phase adopted by the OA·Fe₃O₄ nanoparticles embedded within the nanocomposite fibers, was investigated by X-ray Diffraction (XRD) spectroscopy. The powder XRD pattern of the PEO/PLLA/OA·Fe₃O₄ fibers with 37% and 50% wt magnetic content is provided in Fig. 4. In the same figure the XRD pattern of the pure OA·Fe₃O₄ is also given for comparison. The diffractogram of the fibers display six board peaks appearing at 2θ ∼ 30°, 36°, 43°, 54, 58° and 63°, verifying the presence of magnetite (Fe₃O₄) within the fibers.^53–55

Fig. 4 X-ray Diffraction patterns of the pure OA·Fe₃O₄ nanoparticles and the PEO/PLLA/OA·Fe₃O₄ nanocomposite fibers containing 37% and 50% wt OA·Fe₃O₄.

Thermal and magnetic properties

The fiber decomposition temperatures were determined by TGA. Fig. 5 provides the TGA thermograms of the pure OA·Fe₃O₄, the pristine PEO/PLLA polymer fibers and the nanocomposite fibers containing 37% wt OA·Fe₃O₄. In the case of the pure OA·Fe₃O₄, a weight loss is observed at lower temperatures (below 300 °C), which is attributed to the decomposition of the organic oleic acid coating.⁵¹ The magnetite-free polymer fibers begin to decompose at a temperature lower than ∼300 °C whereas the nanocomposite fibers start to decompose at higher temperature, suggesting that the magnetic nanoparticles affect favorably the thermal stability of the fibers, due to the nanoparticle–polymer interactions.⁵² The remaining residue observed in the case of the magnetite-containing fibers at T > 400 °C, corresponds to the inorganic (Fe₃O₄) content.

Fig. 5 TGA thermograms of OA·Fe₃O₄ nanoparticles, PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ (37% wt OA·Fe₃O₄) electrospun fibers.

The magnetic properties of the Fe₃O₄-loaded nanocomposite fibers were determined by VSM at 300 K. Fig. 6 presents the magnetization versus applied magnetic field strength plots for the two nanocomposite fibers having different magnetic content. The symmetrical sigmoidal shape of the magnetization curves and the lack of a hysteresis loop demonstrate the superparamagnetic behavior of these materials at 300 K. Furthermore, from the obtained magnetization curves it can be concluded that upon increasing the magnetic content within the fibers, the saturation magnetization (Ms) increases, while the superparamagnetic properties are retained.

Fig. 6 Magnetization curves of PEO/PLLA/OA·Fe₃O₄ electrospun nanocomposite fibers containing 37 and 50% wt OA·Fe₃O₄ measured at 300 K.

Surface area measurements

The surface areas determined for the OA·Fe₃O₄-free and the OA·Fe₃O₄-loaded adsorbents were found to be close to the instrument detection limits. However these values were accurate enough to suggest that the materials surface area decreases with increasing magnetite content. This finding is in line with the results obtained from the adsorption studies that are discussed in detail in the following subsection. More precisely, it is demonstrated that the presence of embedded magnetite nanoparticles reduces the fiber adsorption efficiency which is probably attributed to the reduction in the surface area of the fiber in the presence of OA·Fe₃O₄. The surface area values of the PEO/PLLA fibers with different magnetic loading are provided in Table 1.

Table 1 Surface area (S) of the PEO/PLLA fibers containing 0%, 37% and 50% wt OA·Fe₃O₄

Sample	S (m^2 g^−1)
0% wt OA·Fe3O4	1.09 ± 0.09
37% wt OA·Fe3O4	0.51 ± 0.09
50% wt OA·Fe3O4	0.18 ± 0.04

---

Adsorption of MG by the PEO/PLLA and the PEO/PLLA/OA·Fe₃O₄ fibers

An in-depth investigation of the efficacy of the MG removal from aqueous solutions by using the PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ electrospun fibers as adsorbents was carried out, recording to the adsorption isotherms, kinetic experiments and effects of the pH and of the fiber magnetic content. Fig. 7 provides photographs of the PEO/PLLA/OA·Fe₃O₄ electrospun fibers with 50% wt magnetic content prior to and after immersion in MG aqueous solution, attached on a permanent magnet, thus demonstrating the possibility of their magnetic field-assisted recovery from aqueous solutions.

Fig. 7 Photographs of the PEO/PLLA/OA·Fe₃O₄ electrospun fibers (50% wt) prior to and after immersion in MG aqueous solution, attached on a permanent magnet.

FTIR analysis was carried out for the PEO/PLLA and the PEO/PLLA/OA·Fe₃O₄ fibrous mats prior to and after MG adsorption. Fig. 8 shows the FTIR spectra of the pristine PEO/PLLA fibers and of the magnetoactive PEO/PLLA/OA·Fe₃O₄ analogue with 50% wt magnetic content before and after MG adsorption.

Fig. 8 FTIR spectra of PEO/PLLA (0% OA·Fe₃O₄) (a) and PEO/PLLA/OA·Fe₃O₄ (50% wt) (b) fibrous mats recorded prior to and after MG adsorption.

In all FTIR spectra the characteristic absorption band appearing at around 1745 cm⁻¹ corresponds to the backbone ester group of PLLA, whereas the C–O–C stretching bands of PEO appear between 1000 and 1200 cm⁻¹. No significant changes in the FTIR signals can be observed after dye adsorption probably due to the very low concentration of the adsorbed MG molecules. However, a slight shifting of FTIR signals can be seen after MG adsorption, which may be an indication of bond formation between the adsorbent and MG. Similar phenomena (i.e. shifting of FTIR peak values confirming the existence of chemical bonds between the adsorbent and the dye) have been reported by several groups in the past.^58–60

Effect of MG initial solution concentration

In order to evaluate the maximum adsorption capacity (q_max (mg g⁻¹)) of the fibrous mats through the effect of initial MG solution concentration, adsorption experiments with different MG initial solution concentrations have been performed at room temperature (298 K) for 24 hours. The Langmuir isotherm model was employed in order to analyze the data obtained from the adsorption experiments.^61,62 This empirical model assumes that the adsorption takes place at a finite number of homogeneous sites on the surface of the adsorbent until monolayer formation. Once a site is occupied by a molecule, no further adsorption can take place. The Langmuir isotherm model is described by the following equations:where, C_e (mg L⁻¹) is the equilibrium concentration, q_e (mg g⁻¹) is the amount of MG adsorbed at equilibrium per unit mass of adsorbent, q_max (mg g⁻¹) is the maximum adsorption at monolayer coverage and K_d (L mg⁻¹) is the Langmuir adsorption equilibrium constant that reflects the adsorption energy. The UV-Vis spectra of MG aqueous solutions of various initial concentrations, the Langmuir adsorption isotherms expressing the relationship between q_e versus the remaining MG concentration in solution at equilibrium (24 hours), C_e, as well as the linearized 1/q_e versus 1/C_e isotherms corresponding to the MG adsorption by the PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ fibers are provided in Fig. 9(a)–(c) respectively. q_e was determined by using eqn (1).

Fig. 9 UV-Vis spectra of MG aqueous solutions of various initial concentrations (a), q_e vs. C_e Langmuir adsorption isotherms (b), and linearized (1/q_e vs. 1/C_e) Langmuir adsorption isotherms (c), corresponding to the MG adsorption by the PEO/PLLA fibers in the absence and presence of OA·Fe₃O₄ nanoparticles with different magnetic content.

The initial MG concentration C₀ (1 × 10⁻⁶ M to 1 × 10⁻⁵), the solution volume, V (3 mL) and the adsorbent mass W (10 mg) were known, whereas the remaining MG concentration in solution at equilibrium C_e was determined by recording the UV-Vis absorption spectra of the supernatant and calculating the corresponding MG solution concentration by following the Lambert–Beer Law, via the construction of the absorption (at 618 nm) versus MG concentration calibration curve. From the 1/q_e versus 1/C_e plots, q_max and K_d could be calculated from the slope (1/q_maxK_d) and the intercept (1/q_max) (Table 2).

Table 2 Langmuir parameters for MG adsorption determined using electrospun PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ fibers as adsorbents

Sample	qmax (mg g^−1)	Kd (L mg^−1)	R^2
0% wt OA·Fe3O4	8.809	0.399	0.90929
37% wt OA·Fe3O4	1.473	0.376	0.86443
50% wt OA·Fe3O4	1.684	0.404	0.95665

---

The results presented in Table 2 suggest that the presence of magnetite within the fibers reduces the amount of the dye adsorbed. A possible explanation for this phenomenon might be the hydrophobic and cationic nature of the surface of the magnetite nanoparticles at pH values below 8,⁵² that disfavour the incorporation of MG found in its positively charged form (MG⁺) within the fibers, due to the development of electrostatic repulsive forces. Moreover, repulsive phenomena may be also promoted due to the presence of the hydrophobic oleic acid chains employed as steric stabilizers onto the Fe₃O₄ surfaces. Finally, as demonstrated by the surface area measurements, the reduction of the surface area of the fibers upon increasing the magnetic content results to the lowering of the fiber adsorption efficacy.

Adsorption kinetics – effect of contact time

The MG kinetic adsorption profile was investigated upon immersing the PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ fibers in a MG solution of known concentration (C₀ = 1 × 10⁻⁵ M) prepared in de-ionized water and recording the UV-Vis spectrum of the supernatant solution at different time intervals.

Fig. 10 provides a schematic presentation of the MG adsorption process by the fibers and photographs of MG solutions in which the PEO/PLLA (left image) and the PEO/PLLA/OA·Fe₃O₄ (right image) dye adsorbents were immersed. As seen in the photographs, the MG-containing wastewater changed to be almost colorless in the case where the PEO/PLLA fibers were used as adsorbents, whereas fading of the color was also observed when the OA·Fe₃O₄-containing fibers were introduced instead. These results demonstrated that both, the PEO/PLLA and the PEO/PLLA/OA·Fe₃O₄ microfibrous mats could act as effective adsorbents (to a different extent) of MG from water.

Fig. 10 Schematic presentation of the removal of MG by the microfibrous fibers and photographs of the process using the PEO/PLLA (left image) and the PEO/PLLA/OA·Fe₃O₄ (right image) electrospun fibers as dye adsorbents. Photographs were taken 24 hours after fiber immersion in the MG solution.

Fig. 11 provides the absorption of the remaining in solution MG and related adsorption percentage versus time plots corresponding to the adsorption profile of MG from the PEO/PLLA (Fig. 11(a)) and the PEO/PLLA/OA·Fe₃O₄ (Fig. 11(b)) microfibers. Upon immersing the fibers in the MG aqueous solution and recording the UV-Vis spectra of the solution at the indicated times, a systematic decrease in the MG absorption signal corresponding to the unbound MG molecules (appearing at ∼618 nm) was clearly observed. As previously mentioned, by constructing the absorption versus MG concentration calibration curve, it is possible to determine the exact concentration of the non-adsorbed dye at certain time intervals (C_t) and consequently the corresponding adsorption percentage plotted as a function of time in Fig. 11(a) and (b) by using the following equation.

% adsorption = C₀ − C_t/C₀ × 100 (5)

Fig. 11 Kinetic adsorption study of MG from aqueous solutions at room temperature using PEO/PLLA (a) and PEO/PLLA/OA·Fe₃O₄ (b) fibers as adsorbents.

The rates of adsorption were determined by following pseudo-first order kinetics with good correlation. More precisely the ln(q_e − q_t) = f(time) plots were constructed from the slope of which the pseudo-first order rate constants were calculated. Adsorption kinetic constants corresponding to the PEO/PLLA (k = 3.04 × 10⁻² ± 6.079 × 10⁻⁴ min⁻¹) (R² 0.99) and the PEO/PLLA/OA·Fe₃O₄ in the presence of 37% wt OA·Fe₃O₄ (k = 1.08 × 10⁻² ± 2.35 × 10⁻⁴ min⁻¹) (R² 0.99) and 50% wt OA·Fe₃O₄ (k = 1.412 × 10⁻² ± 7.15 × 10⁻⁴ min⁻¹) (R² 0.93) systems reveal slower adsorption rates in the case of the Fe₃O₄-containing fibers whereas lower adsorption percentages were also observed in the presence of magnetite (∼70% compared to ∼95% determined in the absence of Fe₃O₄). Those experimental findings are in line with the aforementioned experimental results related to the maximum adsorption values (q_max) that were found to decrease in the case of the Fe₃O₄-loaded systems.

Fig. 12 summarizes the MG adsorption profiles demonstrated as A_t/A₀ versus time normalized plots for the above-mentioned adsorption studies performed in water. A_t denotes the absorption of the MG molecules found free in solution at time t, and A₀ is the initial absorption (at t = 0) corresponding to the MG molecules found free in solution.

Fig. 12 Normalized A_t/A₀ versus time plots corresponding to the PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ fibers, employed as adsorbents of MG in aqueous solutions at room temperature.

Effect of pH

Among others, pH is a very important parameter which affects the capacity of an adsorbent in wastewater treatment. Specifically in the case of MG, the pH effect on the adsorption process is more pronounced since this particular dye undergoes structural changes depending on the pH.¹⁰ Due to the chromatic changes of MG at different pHs, the UV-Vis spectrophotometry can only be used in the analysis of the MG adsorption efficacy by different adsorbents, based on absorption measurements at the MG characteristic wavelength of 618 nm within a certain pH range (between ∼3 and 6).^6,10 Specifically, in acidic pH (pH < 2) the deep green MG⁺ (having three distinct absorption signals appearing at 317 nm, 427 nm and 618 nm) turns into protonated cyan MGH⁺ whereas in alkaline conditions (pH > 8) it turns into a colorless carbinol base MGOH as seen in Fig. 13. Fig. 14(a) shows the UV-Vis spectra of MG solutions recorded at different pHs prior to fiber immersion whereas in Fig. 14(b) a variation curve of the MG adsorption percentage onto the PEO/PLLA/50% wt OA·Fe₃O₄ fibers as a function of pH, determined upon recording the MG absorption at 618 nm is provided. The effect of pH on MG adsorption by the PEO/PLLA/OA·Fe₃O₄ microfibrous nanocomposite adsorbent was investigated at specific MG concentration (6 × 10⁻⁶ M) and contact time (24 hours).

Fig. 13 Chemical structures of the different MG structures (MGH⁺, MG⁺ and carbinol base) exhibiting colour variations in aqueous solutions, associated with changes in solution pH.

Fig. 14 (a) UV-Vis spectra of MG recorded at different pHs prior to fiber immersion and (b) MG adsorption percentage onto the PEO/PLLA/50% wt OA·Fe₃O₄ fibers as a function of pH, determined upon recording the MG absorption at 618 nm by UV-Vis spectrophotometry.

The reduced adsorption percentage of MG at lower pH values may be attributed to repulsive phenomena developed between the MGH⁺, and the positive charges generated on the magnetite surfaces at pH < 8. Similar observations (i.e. reduced adsorption efficacy for MG at low pH values) have been also reported by other groups^63,64 and have been attributed to two reasons: (a) electrostatic repulsive effects existing between the sorbent and MG owing to the development of positive charges on both, the sorbent and MG under those conditions, inhibiting the MG sorption and (b) competing adsorption effects between the positively charged dye molecules and excess H⁺ ions that are present in solution. At a pH range of 3–6, such phenomena still exist since MG is found in its protonated form (MG⁺) (middle chemical structure, Fig. 13) and magnetite still possesses positive charges on its surface, but they are less pronounced. At pH > 8 the magnetite surface becomes negatively charged and hydrophilic resulting to an increase of the adsorption efficiency of MG that is found in its carbinol base form possessing –OH polar groups.

Adsorption thermodynamics

The influence of temperature on the MG adsorption onto the pristine and nanocomposite electrospun fibers was investigated so as to determine the thermodynamic parameters (ΔH⁰, ΔS⁰ and ΔG⁰) related to the adsorption process. The only variable in the experiments performed was temperature (298, 318, 338 K), whereas the initial MG concentration (1.13 × 10⁻⁵ M) and contact time (24 hours) remained the same. By constructing the ln K_d versus 1/T plot provided in Fig. 15, ΔH⁰ and ΔS⁰ were calculated from the slope (ΔH⁰/R) and the intercept (ΔS⁰/R) respectively according to Van't Hoff eqn (6). K_d is calculated as C_ads/C_e, where C_e (mol L⁻¹) is the equilibrium concentration, C_ads (mol L⁻¹) is the concentration of the amount of MG adsorbed at equilibrium, R (8.314 J K⁻¹ mol⁻¹) is the gas constant and T (K) is the temperature.

ln(K_d) = (ΔS⁰/R) − (ΔH⁰/RT) (6)

Fig. 15 Plots of ln K_d versus 1/T for MG adsorption on the PEO/PLLA and the PEO/PLLA/OA·Fe₃O₄ electrospun fibers.

The Gibbs free energy change ΔG⁰ is calculated from eqn (7).

ΔG⁰ = −RT ln K_d (7)

Table 3 summarizes the calculated values of the thermodynamic parameters ΔG⁰, ΔH⁰ and ΔS⁰. Negative ΔG⁰ values are obtained in all cases, indicating that the adsorption onto the Fe₃O₄-free and the Fe₃O₄-containing fibers is feasible and spontaneous. Moreover, the positive ΔH⁰ values as well as the increase of K_d with increasing temperature⁶⁵ indicate the endothermic nature of the adsorption process. A comparison of the ΔG⁰ values corresponding to the pristine and the Fe₃O₄-loaded adsorbents shows that the adsorption process is thermodynamically less favored in the latter case. This is in line with the above-discussed experimental results that demonstrated the reduced adsorption efficiency of the fibers in the presence of Fe₃O₄ nanoparticles. Moreover, the decrease of the ΔG⁰ values upon temperature increase suggests that the adsorption process is further promoted at higher temperatures.

Table 3 Thermodynamic parameters of MG adsorption on electrospun fibrous mats in the absence and presence of magnetic nanoparticles

Sample code	Temperature (K)	ΔH^0 (kJ mol^−1)	ΔS^0 (kJ mol^−1 K^−1)	ΔG^0 (kJ mol^−1)
0% wt OA·Fe3O4	298	1.0902	0.0134	−2.8785
	318			−3.1447
	338			−3.4112
37% wt OA·Fe3O4	298	0.4986	0.0026	−0.2612
	318			−0.3285
	338			−0.3625
50% wt OA·Fe3O4	298	0.0944	0.0014	−0.3156
	318			−0.3447
	338			−0.3705

---

Desorption/regeneration and recycling studies. The regeneration ability of the fabricated PEO/PLLA and PEO/PLLA/OA·Fe₃O₄ fibers was examined by removing the MG-loaded fibers from the aqueous solution and immersing them in ethanol. An immediate coloration of the alcohol solution was observed, clearly indicating desorption of MG from the fibers and its transfer into the alcohol media. Subsequently, the fibers were re-immersed in a freshly prepared MG solution (of the same initial concentration as that used during the first adsorption step) and the whole desorption/re-adsorption process was repeated for 2 more times. In Fig. 16, the removal efficiency of MG is presented as % adsorption versus number of cycles plot. After three adsorption–desorption cycles, the adsorption capacity was maintained unaffected in both cases, i.e. at ∼88% in the case where PEO/PLLA fibers were used as adsorbents and ∼40% in the case of the Fe₃O₄-containing system, a result that is important for practical applications.

Fig. 16 Regeneration (MG adsorption–desorption) cycles performed using the pristine and the Fe₃O₄-containing microfibers as adsorbents.

Removal of MG from urban wastewater

In an effort to demonstrate the ability of these materials to act as effective adsorbents for MG in real wastewater, MG adsorption experiments were performed in urban wastewater (UWW) samples (secondary treated effluents) by using the pristine PEO/PLLA and the PEO/PLLA/OA·Fe₃O₄ (37% wt) fibers as adsorbents. Upon immersing the fibers in the MG UWW solution and recording the UV-Vis spectra of the solution at different time intervals, a systematic decrease in the MG absorption signal (at ∼618 nm) corresponding to the unbound MG molecules was clearly observed in both cases. Fig. 17 provides the UV-Vis spectra of the UWW supernatant solutions in the case where PEO/PLLA fibers (Fig. 17a) and PEO/PLLA/OA·Fe₃O₄ fibers with 37% wt magnetic content (Fig. 17b) were employed as MG adsorbents, whereas in Fig. 17c the corresponding normalized A_t/A₀ versus time plots are provided. The obtained results verified the efficiency of both systems in removing malachite green oxalate via adsorption from urban wastewater samples, thus demonstrating the potential use of these materials in real wastewater treatment applications.

Fig. 17 MG absorption versus time plots recorded at different time intervals in UWW MG-containing samples at room temperature using PEO/PLLA (a) and PEO/PLLA/OA·Fe₃O₄ (b) fibers as adsorbents and corresponding normalized A_t/A₀ versus time plots (c).

Conclusions

In the present study PEO/PLLA and magnetoactive (superparamagnetic) PEO/PLLA/OA·Fe₃O₄ fibers prepared by means of the electrospinning technique were employed for the first time for the removal of MG from aqueous solutions. The obtained results suggested that the adsorption process follows the Langmuir isotherm model, suggesting monolayer adsorption on a homogeneous surface. Our experimental data show that the presence of Fe₃O₄ nanoparticles in high percentages (∼40–50% wt) within the fibers influences the adsorption process to some extent i.e. it decelerates the adsorption rate and leads to lower adsorption efficiency compared to the pristine polymer fibers. This phenomenon may be attributed to the reduction of the surface area of the fibers upon increasing the magnetic content as suggested by surface area measurements. In addition the obtained results showed that adsorption also depends on operating variables including initial MG concentration and solution pH. The thermodynamic parameters determined from temperature dependent adsorption measurements suggest that the adsorption of MG onto the Fe₃O₄-free and the Fe₃O₄-containing fibers is spontaneous and endothermic. Regeneration/re-use cycles demonstrated the recyclability of these systems. Consequently, the demonstration of the ability of these materials to act as effective adsorbents for MG combined with the possibility of their recovery from aqueous solutions upon applying an external magnetic field and their recyclability, provides a promising tool for the future development of highly efficient, stimuli (magneto)-responsive adsorbents for the removal of hazardous materials from wastewater.

Finally, MG adsorption studies performed in urban wastewater samples by using the PEO/PLLA and the PEO/PLLA/OA·Fe₃O₄ fibers as adsorbents, demonstrated the potential use of these materials in real wastewater treatment applications.

Acknowledgements

This work was supported by the University of Cyprus. We are grateful to Ms. Alina Moaca for the synthesis of the OA·Fe₃O₄ nanoparticles used in the present study. We also thank Prof. Charis Theocharis, Dr P. A. Koutentis (Department of Chemistry, University of Cyprus) and Dr T. Kyratsi (Department of Mechanical and Manufacturing Engineering, University of Cyprus) for providing access to the surface area analysis, TGA/ATR-FTIR and XRD apparatus respectively.

Notes and references

1. G. Jin, G. Pingdao, Y. Li and Z. Fangchuan, J. Hazard. Mater., 2013, 15, 928.
2. M. Tichonovas, E. Krugly, V. Racys, R. Hippler, V. Kauneliene, I. Stasiulaitiene and D. Martuzevicius, Chem. Eng. J., 2013, 229, 9.
3. Z. Aksu, Process Biochem., 2005, 40, 997.
4. A. Witek-Krowiak, Desalin. Water Treat., 2013, 51, 3284.
5. H. R. Rajabi, O. Khani, M. Shamsipur and V. Vatanpour, J. Hazard. Mater., 2013, 250–251, 370.
6. S. Arellano-Cardenas, S. López-Cortez, M. Cornejo-Mazon and J. C. Mares-Gutierrez, Appl. Surf. Sci., 2013, 280, 74.
7. J. Chung and J. O. Kim, Desalin. Water Treat., 2011, 25, 233.
8. P. L. Hariani, M. Faizal, R. Marsi and D. Setiabudidaya, Int. J. Environ. Sci. Dev., 2013, 4, 336.
9. M. Zarei, A. Niaei, D. Salari and A. R. Khataee, J. Electroanal. Chem., 2010, 639, 167.
10. Y. C. Lee, J. Y. Kim and H. J. Shin, Sep. Sci. Technol., 2013, 48, 1093.
11. W. Fan, W. Gao, C. Zhang, W. W. Tjiu, J. Pan and T. Liu, J. Mater. Chem., 2012, 22, 25108.
12. Y. Dong, H. Lin, Q. Jin, L. Li, D. Wang, D. Zhou and F. Qu, J. Mater. Chem. A., 2013, 1, 7391.
13. E. Akar, A. Altinisik and Y. Seki, Ecol. Eng., 2013, 5, 219.
14. M. H. Beak, C. O. Ijagbemi and D. S. Kim, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2009, 44, 536.
15. B. Samiey and A. R. Toosi, Int. J. Chem. Kinet., 2010, 42, 508.
16. T. Santhi and S. Manonmani, Water, 2011, 39, 162.
17. H. Tang, W. Zhou and L. Zhang, J. Hazard. Mater., 2012, 209–210, 218.
18. R. Xu, M. Jia, Y. Zhang and F. Li, Microporous Mesoporous Mater., 2012, 149, 111.
19. N. Sridewi, Y. F. Lee and K. Sudesh, Int. J. Photoenergy, 2011, 2011, 1.
20. E. K. Guechi and O. Hamdaoui, Desalin. Water Treat., 2013, 51, 3371.
21. F. Ferrero and M. Periolatto, Clean Technol. Environ. Policy, 2012, 14, 487.
22. Z. M. Huang, Y. Z. Zhang, M. Kotaki and S. A. Ramakrishna, Compos. Sci. Technol, 2003, 63, 2223.
23. J. Venugopal, Y. Z. Zhang and S. Ramakrishna, J. Nanoeng. Nanosyst., 2005, 218, 35.
24. D. H. Reneker and I. Chun, Nanotechnology, 1996, 7, 216.
25. J. Fang, X. Wang and T. Lin, in Nanofibers-production, properties and functional applications, Intech, 2011, ch. 4, p. 287.
26. S. Laurent and M. Mahmoudi, Int. J. Mol. Epidemiol. Genet., 2011, 2, 367.
27. W. K. Son, J. H. Youk, T. S. Lee and W. H. Park, Macromol. Rapid Commun., 2004, 25, 1632.
28. D. Liang, B. S. Hsiao and B. Chu, Adv. Drug Delivery Rev., 2007, 59, 1392.
29. R. Jayakumar, M. Prabaharan, K. T. Shalumon, K. P. Chennazhi and S. V. Nair, Adv. Polym. Sci., 2012, 246, 263.
30. F. Jian, N. Hai Tao, L. Tong and W. G. Xun Gai, Chin. Sci. Bull., 2008, 53, 2265.
31. S. Cavaliere, S. Subiantoa, I. Savycha, D. J. Jonesa and J. Rozièrea, Energy Environ. Sci., 2011, 4, 4761.
32. S. K. Chang, G. Eun Hee, U. In Chul and H. P. Young, J. Membr. Sci., 2007, 302, 20.
33. S. Yimin, L. Fasheng, G. Qingbao and L. Cunzhen, Desalination, 2008, 223, 349.
34. Y. Tian, M. Wu, R. Liu, Y. Li, D. Wang and J. Tan, Carbohydr. Polym., 2011, 83, 743.
35. I. Savva, M. Efstathiou, T. Krasia-Christoforou and I. Pashalidis, J. Radioanal. Nucl. Chem., 2013, 298, 1991.
36. X. Zhu, X. Jiang, S. Cheng, K. Wang, S. Ma and L. J. Fan, J. Polym. Res., 2010, 17, 769.
37. R. Xu, M. Jia, F. Li, H. Wang, B. Zhang and J. Qiao, Appl. Phys. A: Mater. Sci. Process., 2012, 106, 747.
38. F. Basiri, H. R. S. Abdolkarim, M. Feiz and A. Moheb, Curr. Nanosci., 2011, 7, 633.
39. H. Zhang, H. Nie, D. Yu, C. Wu, Y. Zhang, C. J. Branford White and L. Zhu, Desalination, 2010, 256, 141.
40. F. Xu, X. Shili, S. Mingwu, G. Rui, W. Shanyuan and S. Xiangyang, New J. Chem., 2011, 35, 360.
41. P. Xu, G. M. Zeng, D. Lian Huang, C. L. Feng, S. Hu, M. H. Zhao, C. Lai, Z. Wei, C. Huang, G. X. Xie and Z. F. Liu, Sci. Total Environ., 2012, 424, 1.
42. M. Ayad, N. Salahuddin and A. Fayed, Phys. Chem. Chem. Phys., 2014, 16, 21812.
43. B. Tanhaei, A. Ayati and M. Lahtinen, Chem. Eng. J., 2015, 259, 1.
44. R. R. Mishra, P. Chandran and S. S. Khan, RSC Adv., 2014, 4, 51787.
45. Z. Zhang and J. Kong, J. Hazard. Mater., 2011, 193, 325.
46. I. Savva, A. D. Odysseos, L. Evaggelou, O. Marinica, E. Vasile, L. Vekas, Y. Sarigiannis and T. Krasia-Christoforou, Biomacromolecules, 2013, 14, 4436.
47. D. Bica, Rom. Rep. Phys., 1995, 47, 265.
48. L. Vekas, D. Bica and M. V. Avdeev, China Particuol., 2007, 5, 43.
49. L. Vekas, M. V. Adveev and D. Bica, Magnetic Nanofluids: Synthesis and Structure, in Nanoscience in Biomedicine, ed. D. Shi, Springer, USA, 2009, ch. 25, p. 650.
50. M. V. Avdeev, D. Bica, L. Vekas, V. L. Aksenov, A. V. Feoktystov, O. Marinica, L. Rosta, V. M. Garamus and R. Willumeit, J. Colloid Interface Sci., 2009, 334, 37.
51. B. Samiey and A. Toosi, J. Hazard. Mater., 2010, 184, 739.
52. E. Tombacz, A. Majzik, Z. S. Horvat and E. Ilies, Rom. Rep. Phys., 2006, 58, 281.
53. T. Gong, D. Yang, J. Hu, W. Yang, C. Wang and J. Q. Lu, Colloids Surf., A, 2009, 339, 232.
54. S. Wan, J. Huang, H. Yan and K. J. Liu, Mater. Chem., 2006, 16, 298.
55. W. Jingjing and L. Dengxin, Energy Procedia, 2011, 11, 4794.
56. V. Pérez-Dieste, O. M. Castellini, J. N. Crain, M. A. Eriksson, A. Kirakosian, J.-L. Lin, J. L. McChesney and F. L. Himpsela, Appl. Phys. Lett., 2003, 83, 5053.
57. E. Goiti, M. M. Salinas, G. Arias, D. Puglia, J. M. Kenny and C. Mijangos, Polym. Degrad. Stab., 2007, 92, 2198–2205.
58. M. A. Al-Ghouti, M. A. M. Khraisheh, S. J. Allen and M. N. Ahmad, J. Environ. Manage., 2003, 69, 229.
59. Z. Bekci, Y. Seki and L. Cavas, J. Hazard. Mater., 2009, 161, 1454.
60. C. P. Sekhar, S. Kalidhasan, V. Rajesh and N. Rajesh, Chemisphere, 2009, 77, 842.
61. I. Langmuir, Chem. Rev., 1933, 13, 147.
62. T. Santhi, S. Manonmani and T. Smitha, J. Hazard. Mater., 2010, 179, 178.
63. H. A. Chanzu, J. M. Onyari and P. M. Shiundu, J. Polym. Environ., 2012, 20, 665.
64. M. Anbia and A. Ghaffari, J. Iran. Chem. Soc., 2011, 8, S67.
65. N. Ünlü and M. Ersoz, J. Hazard. Mater., 2006, 136, 272.

---